Title:
|
EFFICIENT DISTRIBUTED REQUEST DISPATCHING FOR LARGE-SCALE CLOUD SERVICES |
Author(s):
|
Po-Jen Shih, Kuo-Chan Huang, Kuan-Chou Lai, Hsi-Ya Chang |
ISBN:
|
978-989-8533-14-2 |
Editors:
|
Hans Weghorn and Pedro IsaĆas |
Year:
|
2012 |
Edition:
|
Single |
Keywords:
|
Cloud services, load balancing, distributed request dispatching |
Type:
|
Full Paper |
First Page:
|
299 |
Last Page:
|
306 |
Language:
|
English |
Cover:
|
|
Full Contents:
|
click to dowload
|
Paper Abstract:
|
Web services on the cloud have to serve a huge amount of users from all over the world simultaneously and usually deploy a large amount of servers to share the incoming workloads. A traditional centralized load balancer cannot scale well under such a large scale of servers. Instead, a set of distributed dispatchers are deployed to achieve the load-balancing goal through appropriate requesting dispatching. The distributed load-balancing structure makes traditional algorithms, such as Join-the-Shortest-Queue (JSQ), inappropriate due to large amounts of communication overheads between dispatchers and servers. A recent research proposed a Join-Idle-Queue (JIQ) algorithm to resolve the above problem for distributed load balancing. However, JIQ is not effective when system load is moderate or high. In this paper, we propose three Join-Queue-Anytime (JQA) based approaches for achieving good load-balancing effects under moderate or high system load. The proposed approaches have been evaluated through a series of simulation experiments implemented with the CloudSim toolkit. The experimental results indicate that the proposed JQA mechanism achieves significant performance improvement compared to the JIQ algorithm, up to 31% improvement in terms of average request response time. |
|
|
|
|